Drivers Of Default Research Articles

Drivers of MSME loan defaults using survival analysis: implications for lenders and policy planners

PurposeThe authors analyse the determinants of loan defaults in micro, small and medium enterprises (MSME) loans in India from the survival duration perspective to draw inferences that have implications for lenders and policymakers.Design/methodology/approachThe authors use the Kaplan–Meier survivor function and the Cox Proportional Hazard model to analyse 4.29 lakhs MSME loan account data originated by a large bank having a national presence from 1st January 2016 to 31st December 2020.FindingsThe estimated Kaplan–Meier survival function by various categories of loan and socio-demographic characteristics reflects heterogeneity and identifies the trigger points for actions. The authors identify the key identified default drivers. The authors find that the subsidy amount is more effective at the lower level and its effectiveness diminishes significantly beyond an optimum level. The simulated values show that the effects of rising interest rates on survival rates vary across industries and types of loans.Practical implicationsThe identified points of inflection in the default dynamics would help banks to initiate actions to prevent loan defaults. The default drivers identified would foster more nuanced lending decisions. The study estimation of the survival rate based on the simulated values of interest rate and subsidy provides insight for policymakers.Originality/valueThis study is the first to investigate default drivers in MSME loans in India using micro-data. The study findings will act as signposts for the planners to guide the direction of the interest rate to be charged by banks in MSME loans, interest subvention and tailoring subsidy levels to foster sustainable growth.

Bankruptcy Prediction using Machine Learning Techniques: Evidence on Indian companies under Insolvency and Bankruptcy Code

This paper attempts to evaluate the predictive ability of four machine learning models: logit, decision tree, random forest and 2-class support vector machine and to identify the key predictors of default. The models are applied on a dataset of 57 companies under the Insolvency and Bankruptcy Code (IBC) in India and a matched sample of 55 solvent companies spanning over ten years from FY06-FY 2016. The solvent companies are matched on size (log of total assets) and sector and are rated ‘AAA’ and ‘AA’. 31 explanatory variables are identified by us for the study which include (i) financial ratios (ii) size and age of the company, (iii) ownership pattern and (iv) market ratios. The empirical findings reveal that random forest strongly outperforms all other models in their predictive ability, followed by SVM, DT and logit model. The findings also confirm relevance of size and age of the firm, market ratios and ownership pattern as predictors of default in addition to financial ratios. We conclude that both parametric (logit) and non-parametric models are useful in the study of default prediction as reflected in the robustness of all models with accuracy of over 75 percent. These models can help banks in strategizing their lending decisions based on credit quality of borrower firms. Our contribution is that to the best of our knowledge this is the first paper that is using the database of companies that are legally defined as insolvent and bankrupt and also taking a balanced sample to avoid biasness and inaccuracy from data imbalance. Also, this study has gone beyond traditional financial statements in identifying key default drivers.

SME CREDIT RISK MODELLING IN SOUTH AFRICA: A CASE STUDY

This paper examined the funding conundrum by assessing the success rate of applications of small and medium sized enterprises (SME) for commercial bank funding. A quantitative group analysis was done on the overdraft obtained from one of the leading financial institutions in South Africa to determine the drivers of default. The SME scorecard was developed using logistic regression on credit applications over a seven-year observation period to analyse the default experience as part of credit risk management. The robustness, stability and relevance of an application scorecard is enhanced by the reject inference process and inclusion of bureau information. Small businesses operating in the service sector and having a long-standing rapport with the bank can easily access commercial bank funding. SMEs in the construction industry with a high number of credit enquiries are unlikely to survive the stringent conditions of the bank lending criteria. It is the prerogative of the principal business owner to honour their financial obligations across the credit industry if commercial bank funding is desired. Their credit quality, as reflected in the bureau information, forms the fulcrum of the SME application scorecard. The model developed in this study can be used as a tool to reduce defaults and serious delinquencies in boarding new applicants. Furthermore, the model can be applied to determine risk tendency and monitor the performance of SME credit portfolios.

Housing loan default in Malaysia: an analytical insight and policy implications

PurposeThe purpose of this paper is to examine the drivers of default in the Malaysian housing market in the light of various policy interventions by the country’s central bank, and the government’s expressed concern to ensure balanced growth in the market. This paper assesses the importance of considering the endogeneity of loan-to-value (LTV) in predicting housing loan default and its implications.Design/methodology/approachIn this paper, the author addresses the endogeneity problem in the LTV variable using two instrumental variables (IV) in this probit regression: national residential property gains tax and the statutory reserve ratio of Bank Negara Malaysia. This study uses the instrumental variable probit model to consider endogeneity bias. This study assumes a latent (unobservable) variable (Y*), representing a borrower’s tendency to default, which is associated linearly with the borrower’s and loan characteristics and other variables (Xi). This study uses individual borrower-level information of 43,156 housing loan borrowers from the files of a well-established housing bank in Malaysia.FindingsThis study’s results confirm that endogeneity causes a substantial difference in the magnitude of the estimated effects of LTV on the default tendency. At the lower values of LTV, the probability of default is over-estimated, and at the higher values, the default probability is substantially underestimated. Endogeneity bias also affects the estimated coefficients of loan and borrower characteristics. The authors find that the interest rate is less relevant in predicting loan default. Other loan characteristics, such as loan age, tenure, payment amount and the built-up area, are relevant. This study’s result confirms that the borrower’s location matters, and an increase in state gross domestic product per capita and an increase in the supply of residential units reduce default probability.Research limitations/implicationsThe present study did not explore the applicability of the “equity theory of default” in the Malaysian housing market. This study did not assess “strategic default” issues and the effect of borrowers’ characteristics, personality traits and self-control of Malaysian housing loan borrowers in the mortgage decision-making process. The evolving dynamics of the Malaysian housing market microstructure in property valuation remained unexplored in the present study.Practical implicationsThe findings have crucial relevance in the decision-making process of commercial banks, the central bank and the government to frame policies to foster balanced growth and development in the housing market. The authors argue that striking a subtle balance between the concerns of financial stability and productive risk-taking by commercial banks in Malaysia remains a continuing challenge for the country’s central bank. The authors also argue that designing suitable taxation policies by the government can deliver its cherished goal of balanced development in the housing market.Originality/valueEmpirical research on the Malaysian housing market based on micro-level data is scarce due to a paucity of relevant data. This study is based on the individual borrower-level information of 43,156 housing loan borrowers from the files of a well-established housing bank in Malaysia. In this analysis, the authors find clear evidence of endogeneity in LTV and argue that any attempts to decipher the default drivers of housing loans without addressing the issue of endogeneity may lead to faulty interpretation. Therefore, this study is unique in recognizing endogeneity and has gone deeper in identifying the default drivers in the Malaysian housing market not addressed by earlier papers.

Loan to value ratio and housing loan default – evidence from microdata in India

PurposeThis study aims to investigate the factors that drive housing loan default in India based on unique micro-level data drawn from a public sector bank's credit files with a national presence in India. The authors address endogeneity in the loan to value ratio (LTV) while deciphering the drivers of default.Design/methodology/approachThe study uses a probit regression approach to analyze the relationship between the probability of default and the explanatory variables. The authors introduce two instrumental variables to address the issue of endogeneity. The authors also add state-level demographic and several other control variables, including an indicator variable that captures the recent regulatory change. The authors’ analysis is based on 102,327 housing loans originated by the bank between January 2001 and December 2017.FindingsThe authors find that addressing the endogeneity issue is essential to specify default drivers, especially LTV, correctly. The nature of employment, gender, socio-religious category and age have a distinct bearing on housing loan defaults. Apart from the LTV ratio, the other key determinants of default are the interest rate, frequency of repayment, prepayment options and the loan period. The findings suggest that the population classification of branch location plays a significant role in loan default. The authors find that an increase in per capita income and an increase in the number of employed people in the state, which reflects borrowers' ability to pay by borrowers, reduce the probability of default. The change in the regulatory classification of loan assets by the Reserve Bank of India did not bear the main results.Research limitations/implicationsThe non-availability of the house price index in analyzing the default dynamics in the Indian housing finance market for the period covered under the study has constrained our analysis. The applicability of the equity theory of default, strategic default, borrowers' characteristics and personality traits are potential research areas in the Indian housing finance market.Practical implicationsThe study's findings are expected to provide valuable inputs to the banks and the housing finance companies to explore and formulate appropriate strategic options in lending to this sector. It has highlighted various vistas of tailor-making housing loan product offerings by the commercial banks to ensure and steady and healthy growth of their loan portfolio. It has also highlighted the regulatory and policy underpinnings to ensure the healthy growth of the Indian housing finance market.Originality/valueThe study provides a fresh perspective on the default drivers in the Indian housing finance market based on micro-level data. In our analysis, the authors find clear evidence of endogeneity in LTV and argue that any attempts to decipher the default drivers of housing loans without addressing the issue of endogeneity may lead to faulty interpretation. Therefore, this study is unique in recognizing endogeneity and has gone deeper in identifying the default drivers in the Indian housing market not addressed by earlier papers on the Indian housing market. The authors also control for the regulatory changes in the Indian housing finance market and include state-level control variables like per capita GDP and the number of workers per thousand to capture the borrowers' ability to pay characteristics.

A Comparison of Borrowing and Default Behaviour Between Men and Women

Micro-credit started out as an intervention designed to address credit scarcity of women, purportedly because women are disenfranchised from property rights and adhere to repayment norms. Men, in contrast, were known for misusing the loans and defaulting on them. This paper, based on a sample of 490 men and 349 women drawn from four districts in erstwhile Andhra Pradesh, compares the borrowing and default drivers of men and women using systematic sampling techniques and rigorous statistical tests. Results show women’s disenfranchisement from family income and property, and hence their dependence on SHG membership for borrowing. Furthermore, men borrow from banks and planned repayments even in conditions of distress, where both genders are cost rationed. The findings also point to men co-opting with women for financing agriculture. A comparison of default drivers points to debt traps and lower or inadequate income driving defaults and underscores the asymmetry of information and asymmetry of power between lender–principals and borrower–agents, where the principals have an advantage over the agents in both the cases ( Saam, 2007 ). The paper, therefore, questions the blanket assumptions of compliant repayments by women and wilful defaults by men. JEL codes: D63, D82, J16, Z13, C51

An innovative feature selection method for support vector machines and its test on the estimation of the credit risk of default

AbstractSupport vector machines (SVM) have been extensively used for classification problems in many areas such as gene, text and image recognition. However, SVM have been rarely used to estimate the probability of default (PD) in credit risk. In this paper, we advocate the application of SVM, rather than the popular logistic regression (LR) method, for the estimation of both corporate and retail PD. Our results indicate that most of the time SVM outperforms LR in terms of classification accuracy for the corporate and retail segments. We propose a new wrapper feature selection based on maximizing the distance of the support vectors from the separating hyperplane and apply it to identify the main PD drivers. We used three datasets to test the PD estimation, containing (1) retail obligors from Germany, (2) corporate obligors from Eastern Europe, and (3) corporate obligors from Poland. Total assets, total liabilities, and sales are identified as frequent default drivers for the corporate datasets, whereas current account status and duration of the current account are frequent default drivers for the retail dataset.

Finding default? Understanding the drivers of default on contracts with farmers’ organizations under the World Food Programme Purchase for progress pilot

AbstractDuring the past two decades, food assistance policy has shifted toward local or regional food purchases and away from purchases from donor countries. Although most local and regional procurement occurs through hard tendering processes open to large‐scale firms and farms, there is growing interest in identifying how and whether to procure from smallholder farmer organizations (FOs). To date, little is known about the drivers of successful contracting with FOs. We utilize data from the United Nations World Food Programme Purchase for Progress pilot in three East African countries to predict defaults on contracts. We examine four possible explanations: country contexts, FO characteristics, prior experience with contracts, and contract modalities and their relationship to local spot market prices. The single most important predictor of default is the increase in market prices between contract approval and delivery. Yet, although increases in market prices are linked to increases in default, this relationship is decreasing in contract size, indicating search costs associated with breaking contracts. Our findings yield both generalizable and context‐specific insights about whether—and when—procuring from smallholder farmers can be successfully integrated into the food assistance toolkit.

Modelling Credit Default in Microfinance—An Indian Case Study

Credit score models have been successfully applied in a traditional credit card industry and by mortgage firms to determine defaulting customer from the non-defaulting customer. In the light of growing competition in the microfinance industry, over-indebtedness and other factors, the industry has come under increased regulatory supervision. Our study provides evidence from a large microfinance institutions (MFI) in India, and we have applied both the credit scoring method and neural network (NN) method and compared the results. In this article, we demonstrate the capability of credit scoring models for an Indian-based microfinance firm in terms of predicting default probability as well the relative importance of each of its associated drivers. A logistic regression model and NN have been used as the predictive analytic tools for sifting the key drivers of default.

A Survival Approach to Prediction of Default Drivers for Indian Listed Companies

The objective of the research study is to identify the key predictors that can explain default risk for Indian listed companies using survival analysis. The author has applied the semi-parametric Cox proportional hazard model to test the impact of financial ratios, capital market ratios, macro-economic variables, size and age of companies, and the ownership structure of promoters to a dataset of 859 companies panning across 10 sectors. Unlike traditional models on default prediction, survival models focus on “time to default” as the dependent variable. The empirical findings reveal that return on capital employed (ROCE), return on net worth (ROE), interest coverage ratio, exchange rate volatility, GDP growth rate, stock index, promoters holdings % and the percent of shares pledged are all significant predictors of default. Among the market variables, it is seen that beta and the ratio of market value of equity/book value of debt are statistically significant variables in explaining default risk. The empirical findings also generate the hazard ratio for each covariate which examines the predicted change in the hazard for a unit increase in the predictor. The author extends the research by applying the market-based KMV structural model to obtain continuous observations of default probability and regressing the same against all the covariates (Gupta et al., 2013) [1]. It is observed that the set of significant covariates are almost common to those generated by our survival approach. The study concludes in emphasizing the significance of survival models in default prediction as unlike traditional accounting-based and market-based models, these models assess relationship between survival time and covariates. The application of survival models is strongly recommended for credit risk evaluation and modeling as structuring of loans can be done by lenders by assessing the survival times of different firms across the entire observation period being considered.

An Empirical Analysis of Default Prediction Models: Evidence from lndian Listed Companies

This paper attempts to evaluate the predictive ability of three default prediction models: the market-based KMV model, the Z-score model using discriminant analysis (DA), and the logit model; and identifies the key default drivers. The research extends prior empirical work by modeling and testing the impact of financial ratios, macro-economic factors, corporate governance and firm-specific variables in predicting default. For the market-based model, the author has extended the works of KMV in developing a suitable algorithm for determining probability of default (PD). While for the KMV model, the continuous observations of PD are used as the dependent variable, for the accounting-based models, ratings assigned are the proxy for default (those rated ’D’ are defaulted and rated ‘AAA’ and ‘A’ are solvent). The research findings largely support the hypothesis that solvency, profitability and liquidity ratios do impact the default risk, but adding other covariates improves the predictive ability of the models. Through this study, the author recommends that accounting –based models and market based models are conceptually different. While market-based models are forward looking and inclusion of market data makes the default risk quantifiable; to make the PD more exhaustive, it is important to factor in the information provided in the financial statements. The conclusions drawn are that the disclosures in financial statements can help predict default risk as financial distress risk is likely to evolve over time and will be reflected in financial statements beyond accounting ratios. Moreover this will also help divulge “creative accounting” practices by corporates.

Start-Up Banks’ Default and the Role of Capital

Regulation requires banks to hold a minimum capital endowment upon their establishment. But what role does initial capital play in a bank’s lifecycle? This paper addresses the issue for start-up banks. We use both survival-time and binary choice models for a sample of newly-established Italian banks in the period 1994-2006, controlling for a broad set of possible drivers of default, such as market, managerial and financial variables. Our results suggest that initial capital does play a leading role in explaining both the timing and the likelihood of a failure. Other important drivers are organisation and a balanced growth path, while market and management variables appear to play a minor role. We then turn to a quasi-experimental design: exploiting a regulatory shift in 1999 we run a counterfactual analysis of the impact of a regulatory tightening of initial capital, which affected only a subsample of banks. The set of results suggests that the effect on banks’ survival may be significant.

Fear, Shame and Guilt: Economic and Behavioral Motivations for Strategic Default

This study examines underwater primary resident homeowners to identify why some decide to strategically default while others do not. We find that realized shame and guilt are consistent with ex ante expectations. However, the financial backlash experienced by strategic defaulters is less than anticipated, causing strategic defaulters not to regret their actions. State‐specific bankruptcy exemption levels and real estate laws only marginally explain the decision to strategically default, partly because the decision to walk away from a mortgage is emotional, and partly because the implementation of these laws is uncertain and confusing to distressed borrowers. Rather, we find key strategic default drivers include the homeowner's expectation of future real estate price movements, frustration with the lender, moral evaluation of the decision to strategically default, loan knowledge, political ideology, gender, income and age.

Relative Indicators of Default Risk Among UK Residential Mortgages

We have assembled a unique loan-level performance dataset for mortgages originated in the UK to study the differences in default likelihood between loans of varying borrower and loan characteristics. We can broadly confirm the relevance of most commonly known riskfactors and find that most drivers of default for prime are also relevant for non-conforming, drivers of repossessions are largely similar to drivers of arrears and information on adverse borrower information dominates any other risk factor. Our study provides many more details and compares results with recent studies for the US and other European countries.